Interval colorectal cancer risk by haemoglobin concentration in Australia's National Bowel Cancer Screening Program

Background Interval colorectal cancers (CRCs) are an important measure of screening performance. We assessed interval CRC risk after negative faecal immunochemical testing (FIT), its association with faecal haemoglobin (f-Hb) concentration, and differences by age and sex in Australia s National Bowel Cancer Screening Program (NBCSP). Methods We conducted a population-based retrospective cohort study linking Queensland screening and cancer registry records for NBCSP participants aged 50-74 years screened in 2018-2022. Eligible episodes had f-Hb below 20 ug/g in both samples. We excluded episodes with CRC diagnosed on or before follow-up start. Follow-up ended at CRC, the next screening episode, death, 24 months, or December 31, 2023. We estimated adjusted cause-specific hazard ratios (HRs), accounting for repeated episodes within participants, and cumulative incidence with death and subsequent screening as competing events. Findings Among 1,083,149 negative screening episodes from 700,734 participants, 586 interval CRCs occurred over 2,056,405 person-years. The incidence rate was 28.5 per 100,000 person-years, and 24-month cumulative incidence was 57.6 per 100,000 episodes (95% CI 53.1-62.5). Risk increased with age and was higher in males (adjusted HR 1.28, 95% CI 1.08-1.50). The complete-case quantitative cohort included 326,931 episodes and 138 cancers. Compared with maximum f-Hb of 0-5 ug/g, concentrations >15-<20 ug/g were associated with an adjusted HR of 14.93 (9.02-24.70). Corresponding 24-month risks were 35.7 (28.5-44.7) and 504.5 (323.4-787.0) per 100,000 episodes. Higher negative-range f-Hb concentrations were more common in older participants and males. The quantitative cohort showed no clear sex difference after adjustment for f-Hb. Interpretation CRC risk varies substantially among people with a negative screening result. Quantitative F-Hb levels may help identify who is at higher risk of interval CRC. Further research is needed to confirm these findings and assess the benefits, harms and costs of changing screening intervals.